Replacement of Several Single Amino Acid Side Chains Exposed to the Inside of the ATP-binding Pocket Induces Different Extents of Affinity Change in the High and Low Affinity ATP-binding Sites of Rat Na/K-ATPase*

To investigate the relationship between the high and the low affinity ATP-binding site, which appears during the Na+/K+-ATPase reaction, four amino acids were mutated, the side chains of which are exposed to inside of the ATP-binding pocket. Six mutants, F475Y, K480A, K480E, K501A, K501E, and R544A, where the numbers correspond to the pig Na+/K+-ATPase α-chain, were expressed in HeLa cells. The apparent affinities were determined by high affinity ATP-dependent phosphorylation and by the low affinity activation of Na+/K+-ATPase or low affinity ATP inhibition of K+-para-nitrophenylphosphatase (pNPPase). For the mutants K480A and K501A, little affinity change was detected for either the high affinity or the low affinity effect. In contrast, the other four mutants reduced both apparent affinities. Strikingly, R544A had a 30-fold greater effect on the high affinity ATP site than the low affinity site. For the F475Y mutant, it is likely that there was a greater effect on the low affinity site than the high affinity site, but for both F475Y and K480E the affinity for the low affinity ATP effect was reduced so much that it was not possible to estimate a K 0.5. However, both the affinities for the K480E were reduced to ∼1/20. The turnover number of the Na+/K+-ATPase and the apparent affinity for Na+ and pNPP was reduced slightly or not at all for these mutants, but the turnover number of K+-pNPPase and the apparent affinity for K+ were increased. These and other data suggest the presence of only one ATP-binding site, which can change its conformation to accept ATP with a high and low affinity. The requirement of Arg-544 and possibly Lys-501 is more important in forming a high affinity ATP binding conformation, and Phe-475 and possibly Lys-480 are more important in forming the low affinity ATP binding conformation.

exposed to the ATP-binding pocket affects both affinity sites for direct ATP binding with a high and low affinity or whether it affects either affinity site. It is important to understand the relationship between high and low affinity ATP binding for developing a better understanding of the mechanism by which Na ϩ /K ϩ -ATPase operates.
To investigate these points, 10 mutants carrying a single amino acid substitution of the side chain that possibly serve as ligands for ATP in the ATP-binding pocket were constructed. To aid the reader in following these experiments, models of the ATP-binding pocket of the wild type and mutants are presented (Scheme I). In this scheme, the three-dimensional structure of the ATP-binding pocket of the wild type and mutants of Na ϩ / K ϩ -ATPase was fitted to Ca 2ϩ /H ϩ -ATPase (8) by Swiss-Model (www.expasy.ch/spdbv/). The Phe-475 (green), Lys-480 (orange), Lys-501 (yellow), and Arg-544 (violet) and mutated residues are shown as a space-filling model, where the number of the residues corresponds to the pig Na ϩ /K ϩ -ATPase ␣-chain. In this study, where a single mutation induced an apparent affinity change in both the high and low affinity ATP effects, Na ϩ , K ϩ , and pNPP were estimated. The data show that each single amino acid substitution, such as F475Y, K480E, K501E, and R544A, reduced the apparent affinity of the high and low affinity ATP effects to different extents. The data suggest that the Arg-544 and possibly Lys-501 are more important for high affinity ATP binding, whereas the Phe-475 and possibly Lys-480 are more important for low affinity ATP binding. This is the first molecular biological approach to show the importance of an exposed amino acid side chain ligated to ATP for the high and low affinity binding in the pocket. In other words, the ATP-binding pocket changes its conformational state to reflect the high and low affinity ATP binding that accompanies ATP hydrolysis. These data are consistent with the oligomeric nature of the enzyme (5-7), as obtained by molecular weight estimation (21) as well as electron microscopic observation (22), ligand binding stoichiometry during ATP hydrolysis (7,22), cross-linking experiments (6,23), the binding of suicide substrates (5), and restraint-based comparative modeling (20).

MATERIALS AND METHODS
Plasmids-pGEM-NaK and pCDL-NaK containing the entire coding region of rat Na ϩ /K ϩ -ATPase ␣1 cDNA has been described previously (24). The DNA fragments encoding the rat Na ϩ /K ϩ -ATPase ␤ subunit were amplified by the PCR using a pair of sense and antisense oligo-nucleotide primers: for sense, 5Ј-AGCACTCGCTTTCCCTC-3Ј, corresponding to the nucleotides 413-429; for antisense, 5Ј-GGTCCCAT-ACGTATGAC-3Ј, corresponding to the nucleotides 1455-1471 (25). The PCR products were inserted in a pCRII vector (Invitrogen) and transferred to the mammalian expression vector pCDL (pCDL-NaK␤). The sequence of the cDNA insert was confirmed by the dideoxy-mediated chain termination method (26).
HeLa Cells Expressed Rat Na ϩ /K ϩ -ATPase ␤1 Subunit-HeLa cells were cotransfected with plasmid DNAs containing a 10:1 molar ratio of pCDL-NaK␤ to the pKNH vector (28) carrying neomycin-resistant gene by the calcium phosphate precipitation method (29) with an CellPhect Transfection Kit (Amersham Biosciences), selected in 500 g/ml G418, and several cell lines were isolated. The total RNAs were isolated from each cell lines and treated with DNase I. A pair of PCR primers, 5Ј-GGA AGA AAT TCA TCT GG-3Ј and 5Ј-GGTCCCATACGTATGAC 3Ј corresponding to the nucleotides 494 -510 and 1455-1471 of rat ␤ subunit cDNA were designed (24). These regions were the same DNA sequences found in both the rat and human ␤ subunits (30). By using the 32 Plabeled sense primer, the 978-or 982-bp DNA fragment from rat or human cDNA was amplified and was found to contain three or four recognition sites for restriction enzyme PstI, respectively. After PstI treatment, the 692-or 415-bp DNA fragment derived from rat or human cDNA was radioactive. The ratio of the radioactivity of the 692-bp fragment to that of the 415-bp fragment indicates the expression level of rat Na ϩ /K ϩ -ATPase ␤ subunit to the human one. The selected cell line (A6 cell line; the ratio was 1.23) expressed similar amounts of mRNA for the Na ϩ /K ϩ -ATPase ␤ subunit to that for the endogenous ␤ subunit.
Transfection of HeLa Cells Expressing Rat ␤ Subunit with Rat Na ϩ / K ϩ -ATPase ␣1 Subunit-HeLa cells expressing the rat ␤ subunit (A6 cell line) were cultured in a 35-mm dish in Dulbecco's modified Eagle's medium containing 10% fetal calf serum. Cells were transfected with 2 g of plasmid DNA containing a wild type or mutant Na ϩ /K ϩ -ATPase ␣1 subunit cDNA with the GenePORTER transfection reagent (Gene Therapy Systems) and subjected to selection in 10 M ouabain. Ouabain-resistant cells were expanded into stable cell lines.
Preparation of SDS-purified Membrane Vesicles-Crude plasma membranes were prepared from HeLa cells (24) or pig kidney microsomes essentially as described previously (31). Crude plasma membranes at 1 mg/ml were incubated with 0.15 mg/ml of SDS in a buffer containing 0.25 M sucrose, 10 mM dithiothreitol, and 10 mM Tris-HCl, pH 7.4, for 5 min at room temperature. The sample was loaded onto stepwise sucrose density gradients consisting of 10 and 40% sucrose layers and centrifuged at 350,000 ϫ g in a Beckman TLA-100.3 rotor for 10 min at 4°C. The fractions at the 10:40% interface were collected and pooled. The pooled sample was diluted with 3 volumes of the sucrose buffer and centrifuged at 350,000 ϫ g in a Beckman TLA-100.3 rotor for 10 min. The pellet was suspended in the sucrose buffer, and the protein concentration was estimated (32) with bovine serum albumin as a standard.
Phosphoenzyme (EP) Formed from ATP-SDS-treated membrane vesicles (20 g) were incubated at 0°C for 10 s in 100 l of 16 mM NaCl, 0.43 mM MgCl 2 , 25 mM imidazole HCl, pH 7.2, and various concentrations of [␥-32 P]ATP. The reaction was stopped by adding 500 l of an ice-cold 10% trichloroacetic acid solution containing 10 mM inorganic phosphate and 1 mM ATP. The samples were centrifuged at 15,000 ϫ g for 10 min at 4°C. The precipitates were washed with ice-cold 10% trichloroacetic acid solution by centrifugation. The resulting precipitates were washed with ice-cold water by centrifugation. The precipitates were dissolved in the sample buffer containing a trace amount of a bromphenol blue, 10% glycerol, 1% SDS, 5% ␤-mercaptoethanol, and 10 mM sodium phosphate buffer, pH 6.0, and subjected to SDS-PAGE at pH 6.0 (33). After drying the gels, the radioactivity of 32 P incorporated into the Na ϩ /K ϩ -ATPase ␣ subunit was detected and quantitated with a BAS 2000 system (Fuji). In order to correct for losses due to hydrolysis SCHEME 1 of EP during the above procedure, the SDS-treated Na ϩ /K ϩ -ATPase preparation from pig kidney (20 g) was incubated under the same conditions as described above. The final precipitates were dissolved in the sample buffer in the absence of bromphenol blue, and half of the aliquots was taken for estimation of the amount of EP by liquid scintillation counting, and the other aliquot (0.75 g of protein) was subjected to SDS-PAGE and quantitated with a BAS system. The data were analyzed using nonlinear least squares regression (GraphPad Prism; GraphPad Software Inc.) by using the Michaelis-Menten equation: ) in which EP max is the maximum amount of EP formed, and K 0.5 h is the concentration of ATP giving the half-maximum formation of EP.
Na ϩ /K ϩ -ATPase Activity-The ATPase activity of SDS-treated membrane vesicles from HeLa cells was measured at 37°C for 30 min in a reaction medium containing 0.5-2 g of enzyme protein, various concentrations of ATP, 40 mM NaCl, 16 mM KCl, 5 mM MgCl 2 , 125 mM sucrose, 0.5 mM EDTA, and 20 mM Tris-HCl, pH 7.4, in the presence of 5 M or 5 mM ouabain. The reactions were stopped by adding an equal volume of 12% SDS. The colorimetric determination of inorganic phosphate with ammonium molybdate complexes was performed (34). Rat Na ϩ /K ϩ -ATPase activity was determined as the difference between the activities in the presence of 5 M and 5 mM ouabain. The activity was divided by the amount of EP max estimated as described above and expressed as s Ϫ1 . All determinations were performed in duplicate. The data were analyzed in essentially the same manner as described above, where EP and EP max were, respectively, replaced with ATPase activity (s Ϫ1 ) and V max (the maximum turnover number). Dependence on Na ϩ or K ϩ concentration of Na ϩ /K ϩ -ATPase activity was measured with 10 M or 5 mM ouabain at 37°C for 30 min in the presence of 5 mM MgCl 2 , 4 mM ATP, and various concentrations of NaCl or KCl. The data were analyzed by nonlinear least squares regression by using the Hill equa- is the maximum activity, n H is the Hill coefficient, and K 0.5 is the concentration of ion giving half-maximum ATPase activity.
Potassium-dependent pNPPase Activity-The pNPPase activity of the SDS-treated membrane vesicles was measured in duplicate in the presence or absence of 16 mM KCl at 37°C for 30 min containing 1-2 g of enzyme protein, various concentrations of pNPP, 6 mM MgCl 2 , 125 mM sucrose, 0.5 mM EDTA, and 20 mM imidazole HCl (pH 7.2). The K ϩ -pNPPase activity was the difference between the activity with and without KCl. The K ϩ -pNPPase activity was also measured in the presence of 5 mM pNPP and various concentrations of ATP to estimate the K i,0.5 for ATP by nonlinear least squares regression.

Construction of Stable Cell Lines
Expressing the Rat Na ϩ / K ϩ -ATPase ␣ Subunit Variant in HeLa Cells-The 10 expression plasmid vectors (pCDL-NaKs) carrying a single amino acid substitution at Phe-475, Lys-480, Lys-501, or Arg-544, namely F475A/F475S/F475D/F475Y, K480A/K480E, K501A/K501E, and R544A/R544E, were constructed and transfected to HeLa cells expressing the rat Na ϩ /K ϩ -ATPase ␤ subunit. HeLa cells transfected with the six pCDL-NaKs containing rat ␣1 cDNAs mutated to encode the amino acid substitutions F475Y, K480A/ K480E, K501A/K501E, and R544A resulted in the appearance of HeLa cells that survived in a medium containing 10 M ouabain. This condition permitted cells to be obtained with a higher level of rat Na ϩ /K ϩ -ATPase activity than before (24) without further steps of cell cloning. Ouabain-resistant cell lines by transfecting with the four pCDL-NaKs encoding the F475A/F475S/F475D or R544E could not be obtained, which suggests that these mutations impaired the Na ϩ /K ϩ -ATPase activity to an extent that the cells did not survive.
Effect of Each Substitution of Phe-475, Lys-480, Lys-501, and Arg-544 on the Apparent High Affinity ATP Binding Registered by EP Formation-To estimate the apparent affinity for the high affinity ATP effect, the amount of ATP-dependent EP formation in an SDS-treated enzyme preparation from pig kidney microsomes was measured first in the presence of 0.43 mM Mg 2ϩ , 2 M ATP, and 16 mM NaCl (Fig. 1A, left panel). The amount of EP detected in the membranes increased linearly with increasing amounts of added enzyme protein (not shown). Fig. 1A (right panel) shows that the phosphorylation of SDS-treated HeLa cell membranes expressing the rat wild type Na ϩ /K ϩ -ATPase was Na ϩ -dependent. The amount of steady state EP for the SDS-treated membrane of the rat wild type and mutants was measured in the presence of different concentrations of ATP. The data followed Michaelis-Menten-type kinetics ( Fig. 1, B and C), which gave the maximum amount of phosphoenzyme (EP max ) and a half-maximum concentration of ATP (K 0.5 h ). The K 0.5 h of the wild type and the K480A were similar and low in value, a mean value ϳ0.026 M, and that of K480E, K501A, and K501E increased, respectively, to 19-, 2-, and 9-fold. The value of F475Y and R544A increased to 5-and 59-fold (Table I,  Effect of the Substitution on the Apparent Low Affinity ATP Binding Registered by Na ϩ /K ϩ -ATPase Activity and ATP-induced Inhibition of K ϩ -pNPPase Activity-To estimate the apparent affinity for the low affinity ATP effect, Na ϩ /K ϩ -ATPase activities of the rat wild type and mutants were measured with increasing ATP concentrations up to 4 mM in the presence of 40 mM NaCl, 16 mM KCl, and 5 mM MgCl 2 and either 5 M or 5 mM ouabain. The difference between these activities was assumed to be the Na ϩ /K ϩ -ATPase activity (V) of expressed enzyme. The data were fitted to the Michaelis-Menten equation (Fig. 2), which gave the half-maximum concentration of ATP (K 0.5 l ) and the V max as shown as 100%. The turnover number, V max /EP max , is shown in Fig. 2, where the value of EP max was obtained in previous section. The activities of the F475Y and K480E were essentially linear in the ATP concentration range studied. Thus the activity in the presence of 4 mM ATP was assumed to be 100% as shown in parentheses. The K 0.5 l of the wild type, K480A, and K501A again showed a similar value, a mean value ϳ0.69 mM, and that of K501E and R544A increased ϳ2-fold (Table I, line b). The K 0.5 l of F475Y and K480E could be estimated as a minimum value, because the activity increased linearly in the presence of 0.1-4 mM ATP. These data suggest that they are greater than 4 mM, up to a ϳ6-fold increase. The change in the V max /EP max is rather small, around a 50% reduction for the case of K501A/K501E and R544A (Table I, line c).
It is well known that a low affinity ATP binding to the enzyme inhibits K ϩ -pNPPase activity (35). Thus, the apparent affinity (K i,0.5 ) for the low affinity ATP binding was estimated by measuring ATP-induced inhibition of the K ϩ -pNPPase activity (Fig. 3). The K i,0.5 of the wild type and K501A gave a similar value, a mean value ϳ0.36 mM. Each replacement of Ala with Glu in Lys-480 and Lys-501 increased the K i,0.5 , respectively, to 21-and 7-fold and the value of F475Y and R544A increased by 8-and 4-fold, respectively (Table I, line d).
Effect of the Substitution on the Apparent Affinity for pNPP Registered by K ϩ -pNPPase Activity-The hydrolysis of pNPP in the presence of K ϩ and Mg 2ϩ of Na ϩ /K ϩ -ATPase was assumed to occur in the K ϩ -occluded enzyme form (36) without the formation of E 1 P and E 2 P. Mutants such as F475Y, K480E, K501E, and R544A showed a reduced affinity in both or either ATP binding (Table I, lines a-c). To investigate the relationship between a mutation in the ATP-binding pocket and K ϩ -pNPP hydrolysis, pNPPase activity was measured in the presence of 6 mM MgCl 2 and various concentrations of pNPP with and without 16 mM KCl. The K ϩ -pNPPase activity (V) represents the difference in the activity in the presence and absence, respectively, of 16 mM KCl. The addition of 1 mM ouabain reduced the activity to the level found in the absence of K ϩ and suggests that the K ϩ -pNPPase activity is due to a property unique to Na ϩ /K ϩ -ATPase. The data were fitted to Michaelis-Menten kinetics. Fig. 4 shows the plots of V against the concentration of pNPP, where V max p as 100% and the turnover number, V max p /EP max , are also shown (Fig. 4). In this case, the value of EP max was taken from data in Fig.  1. Both the turnover number, V max p /EP max , and the K 0.5 p  To compare the activity between the wild type and mutants, the ratio (V max to EP max) estimated as described in Fig. 1  were altered only slightly as a result of these mutations, i.e. the maximum changes were ϳ2-fold increases in R544A (Table I, lines e and f).
Effect of the Substitution on the Apparent Affinity for Na ϩ and K ϩ Registered by Na ϩ /K ϩ -ATPase Activity-To investigate the relationship between these mutations and the apparent affinity change for Na ϩ or K ϩ , the Na ϩ /K ϩ -ATPase activity was measured in the presence of different concentrations of Na ϩ or K ϩ as described above except that 4 mM ATP was used. The values K 0.5 for Na ϩ (K 0.5 Na ) and the Hill coefficient (n H ) were estimated using nonlinear least square regression (Fig. 5). The K 0.5 Na for the wild type was 7.5 mM, and the maximum decrease and increase was, respectively, observed in K480A (5.5 mM) and both K501E and R544A (10.5 mM). The data show that a change in the apparent affinity for Na ϩ , if present, is rather small compared with the affinity change for the high and low affinity ATP-binding sites (Table I, compare line g with lines a, b, and d). The n H of the wild type was ϳ1.8 and that of K480A and R544A was 1.3 and 2.3, respectively, which represents the minimum and the maximum value (dotted lines in Fig. 5). The reasons for the decrease or the increase of n H in K480A or R544A, respectively, were not clear.
The dependence of ATPase activity on K ϩ concentration ( Fig.  6 and Table I, line h) showed that the K 0.5 K for the wild type was 0.67 mM, which was the maximum, and all mutants had lower values. These data indicate that these mutants had a slightly increased apparent affinity for K ϩ , i.e. the maximum increase was observed in R544A, namely 2.5 (1/0.4)-fold. The n H of wild type and mutants was ϳ1.5 except for the value of 1.9 for R544A.

DISCUSSION
Flexibility of the ATP-binding Pocket-Each mutant, F475Y, K480E, K501E, and R544A, respectively, showed different extents of apparent decreased affinity between the high and low affinity ATP effects. Two mutants in which the cationic Lys side chain was replaced by a less bulky methyl group (K480A and K501A) had only a slight affinity changes for both ATPbinding sites (Scheme I).
The replacement of Phe-475 with Tyr, namely the addition of a phenolic hydroxyl group, reduced the apparent affinity for the high and low affinity ATP effects, respectively, to 1/5 and Ͻ1/6 or 1/8 without affecting the binding of the less bulky pNPP (Scheme I). Stable cell lines could not be obtained when Phe was replaced with Ala, Ser, and Asp, respectively.
Each replacement of a cationic side chain with a less bulky anionic side chain (K480E and K501E) showed a decreased affinity for both ATP-binding sites. Such a reduction did not occur in the case of replacement with a methyl side chain (K501A).
The replacement of a guanidinium moiety with a methyl side chain (R544A) showed a very strong reduction in affinity (1/59) for the high affinity ATP effect with a slight reduction in affinity for the low affinity ATP effect as well as pNPP binding (1/2) (Table I, lines a, b, and e). The replacement of the Arg with Glu side chain resulted in no appearance of the HeLa cells surviving in the presence of 10 M ouabain. The importance of Arg-544 was also reflected by the fact that R544K and R544Q, respectively, showed only a small effect on the apparent affin- , K501E (f), and R544A () was measured as described in the Fig. 3 except ATP was not included, and various concentrations of pNPP were present. The data were subjected to a nonlinear least square regression, where 100% of K ϩ -pNPPase activity was the value of V max estimated from the fitting. To compare the activity between the wild type and mutants, the ratio V max to EP max estimated as described in Fig. 1 was taken, and the V max /EP max values are shown with the K 0.5 p for pNPP. ity for MgATP with a 30% of Na ϩ /K ϩ -ATPase activity and a 1/30 reduction in the binding of MgATP with little activity (37). Thus the requirement of the Arg side chain (Scheme I) for the high affinity site is considerably more strict than the low affinity site (Table I, lines a, b, and d).
The F475Y, K480A/K480E, K501A/K501E, or R544A showed a slight increase in affinity for K ϩ with a small decrease in affinity for pNPP and Na ϩ (Table I, lines h, e, and g). When the ratios, K K /K Na , were calculated, the increase in K ϩ affinity becomes more clear (Table I, line i) such as the case of K480E, K501E, and R544A, which showed reduced affinity for both ATP effects (Table I, lines a, b, and d). Such an apparent antagonism between ATP binding and K ϩ binding has been also reported for mutations in the cation-binding pocket of Na ϩ /K ϩ -ATPase which showed a decreased affinity for K ϩ and an increase in the apparent affinity for ATP (38 -40).
The present data suggest the importance of the Phe-475, Lys-480, Lys-501, and Arg-544 in the construction of a flexible ATP-binding site that is capable of accepting ATP with a high and low affinity. These side chains appear to be close to the aromatic ring of ATP, interacting with the ␤-phosphate group of ATP, at a distance of 5 Å to the adenosine, and close to the interface to the phosphoryl domain, respectively (20). Flexibility of the binding pocket is indicated from the covalent binding of a rather bulky fluorescence probe such as pyridoxal 5Јphosphate or pyridoxal 5Ј-diphospho-5Ј-adenosine into Lys-480 and/or fluorescein 5Ј-isothiocyanate into Lys-501 (12,15). These modifications had only a small influence on the V max of K ϩ -pNPPase, high affinity ATP binding without EP formation, low affinity ATP binding, and the Na ϩ -dependent E 1 P and E 2 P formation from acetyl phosphate or pNPP, but inhibited Na ϩ / K ϩ -ATPase activity and ATP-dependent EP formation almost completely (12)(13)(14)(15)(41)(42)(43). These data suggest that ATP is able to enter the ATP-binding pocket despite the presence of a bulky probe at the entrance and/or the depth of the ATP-binding pocket (8,20).
Different Contributions of the Exposed Side Chain for ATP Binding-Two different ATP bindings induce different conformational change during turnover (12-15, 35, 41-43), which suggests that the contribution of each side chain in the pocket in inducing two different binding conformations may be different. To investigate this point further, the ratios K 0.5 l /K .5 h and K i,0.5 /K 0.5 h were calculated (Table I, lines j and k). The value of the wild type enzyme was assumed to be 1, and the relative value was compared with those of the mutants. If the value becomes Ͻ1, this suggests that the mutation induced a larger affinity change at the high affinity ATP effect rather than the low affinity ATP effect. Although R544A showed a decrease in both affinities, the extent of the high affinity effect decreased more than that for low affinity effect, namely 30-(ϭ 1/0.03) or 14 (ϭ 1/0.07)-fold, whereas F475Y showed a Ͻ1.2 or 1.7-fold decrease for the low affinity effect compared with that for the high affinity. To compare the affinity ratio for the low affinity ATP binding to the pNPP binding, both of which are assumed to occur in the E 2 state, similar ratios (K 0.5 l /K 0.5 p and K i,0.5 /K 0.5 p ) were also calculated. The data suggest that F475Y and K480E showed a larger reduction in the apparent low affinity ATP binding (Table I, lines l and m) compared with that for pNPP binding. When the affinity ratio for the high affinity ATP binding to the low affinity pNPP binding, K 0.5 h /K 0.5 p , was calculated, R544A showed the largest reduction in apparent high affinity ATP binding compared with that of pNPP (Table I, line n).
These data suggest that the contribution of each side chain to the low and high affinity ATP binding is different. The Arg-544 is more important in achieving the high affinity ATP binding, whereas the Phe-475 is more important in achieving a the low affinity ATP binding.
Are the High and the Low Affinity ATP-binding Sites Present Simultaneously in the Same ␣-Chain?-The present experiments could not answer this question directly, because no quantitative ATP binding measurements were carried out due to the rather low purity (Fig. 1C, inset, 7-17 pmol of EP/mg protein) of the SDS-treated enzyme from HeLa cells. However, by using purified pig kidney enzymes, the maximum 32 P binding/␣-chain during turnover in the presence of [␥-32 P]ATP was found to be ϳ1 mol/mol of ␣-chain due to the sum of E 32 P enzyme-bound 32 P and EAT 32 P (enzyme-bound [␥-32 P]ATP) but not 2 mol/mol of ␣-chain (22). Furthermore, the maximum number of ligand binding/␣-chain, under non-turnover conditions, has been reported to be 1 ATP or 1 ATP analogue, 3 Na ϩ , 2 Rb ϩ , and 1 ouabain as reviewed recently (5). Thus, the maximum binding of 1 mol of 32 P or [␥-32 P]ATP binding/␣chain excludes the possibility of a presence of another hypothetical low affinity ATP-binding site in the same ␣-chain (19,44), the conformational state of which might be affected indirectly by a high affinity ATP binding to the ATP-binding pocket. The presence of a single ATP-binding site, based on restraint-based comparative modeling of the H 4 -H 5 loop of Na ϩ /K ϩ -ATPase (20) and on the atomic structure of Ca 2ϩ /H ϩ -ATPase (8), has been also reported. It was also shown unequivocally that a similar reduction in ATP binding capacity occurred during the turnover of pig gastric H/K-ATPase, from 1 mol/mol of ␣-chain to the half to form EP:EATP and the resulting liberation of each 0.5 mol of P i (45). In other words, ATP binding capacity decreases accompanying the turnover in both Na/K-and H/K-ATPase (22,45).
Each single mutation, such as F475Y, K480E, K501E, and R544A, strongly reduced both the apparent ATP affinities as in the case of a single mutation of Phe-487 (corresponding to Phe-475 of the pig kidney enzyme), Arg-489, and Lys-492 in Ca 2ϩ /H ϩ -ATPase, which has an ATP-binding pocket similar to Na ϩ /K ϩ -ATPase (18). It had been already reported that the modification of Lys-501 of Na ϩ /K ϩ -ATPase with N-(2-nitro-4isothiocyanophenyl)-imidazole affected both the high and low affinity ATP effects (46). These considerations suggest that each ␣-chain contains only one ATP-binding pocket. One accepts ATP with a high affinity to form EP and the other, in the adjacent ␣-chain, accepts ATP with a low affinity to accelerate Na ϩ /K ϩ -ATPase activity. The present finding is consistent with the view that Na ϩ /K ϩ -ATPase functions out of phase (5-7, 21, 23, 47, 48) as a diprotomer, (␣␤) 2 , or a much higher oligomer, (␣␤) 4 , as in the case of cross-talking gastric H ϩ /K ϩ -ATPase (45) and possibly Ca 2ϩ / H ϩ -ATPase (49,50).